When two horses are too close to separate
Two handicaps, same afternoon, favourites at the same price. On the screen they are twins. Underneath they are not remotely the same race.
Two races, one price
In the first, the top of the race is clean. One horse stands clearly above the rest on our internal rating, there is daylight behind it, and the rest of the field arranges itself into a sensible queue. You can say what this race is about in a sentence.
In the second, the top is a scrum. Three or four horses sit within touching distance of each other, and worse, they are close for different reasons. One is there on recent improvement, one on a long consistent record, one because it has been dropped in grade. Their profiles are not just similar in value, they are conflicting in kind. The order at the top would change if you nudged almost any input.
Both favourites are the same price. Everything a punter sees says these are equivalent situations. Reading the two races tells you they are nothing of the sort.
We call the gap between the top of a race and everything under it separation. It is not a rating of the horse. It is a description of the race, and specifically of how much you would have to disturb the evidence before the order at the top changed. Wide separation means the answer is stable. Compressed means it is balanced on something thin.
The obvious next question is whether that difference is worth anything.
Separation does something real
It does. This is the encouraging half.
When the internal rating strongly agrees about a race, meaning the top is genuinely separated rather than compressed, favourites re-sort. The favourites in well-separated races are a measurably better group of horses than the favourites in compressed ones. Their strike rates differ, and the difference is not a rounding record you would talk yourself into after the fact.
So the intuition every form student has, that some favourites are standing on much firmer ground than others of the same price, is not folklore. It survives contact with a large sample. That was worth establishing on its own, because plenty of racing's favourite intuitions do not survive it, and because it means separation is measuring something about the world rather than something about our own arithmetic.
At which point you can feel the idea starting to look like a system. Wait for the well-separated races, back the favourites in those, ignore the scrums. It is neat, it is easy to explain, and I would have been very pleased with it.
The market had already done it
Then you hold the price constant and the whole thing goes quiet.
Observed results against what the prices implied sat close to 1.00 in both groups. The better group won more often, and it was priced to win more often. The weaker group won less often, and it was priced to win less often. The re-sorting was real, and the market had already carried it out before we got there.
This is a pattern I have stopped being surprised by. Real effect, no edge. The two are not the same statement and the gap between them is where a great many racing systems quietly die. You find something true, you get excited, and then you notice the price already knew.
It leaves separation exactly where it should be. It is a description of how legible a race is, which is genuinely useful for deciding whether a race is worth reading at all. It is not a reason to back the favourite, and we have never presented it as one.
Real effect, no edge. Both halves of that sentence matter.
Check today’s races
See how Elite Pass reads today’s eligible races.
Related research
More notes
- The horse has not changed, the conditions haveA horse that ran badly last time usually did not run badly. It was asked a different question, and it answered that one honestly.
- What a rating tells you, and what it cannotA rating is about the horse. A price is about the race. Keeping those two apart turns out to be most of the work.
- Reading the race, not the horseIf the best horse loses two times in three, which one is best is rarely the useful question. Here is the one we ask instead, and how we ask it.
Elite Pass publishes research, not guaranteed outcomes. Findings are measured against the market and remain subject to replication and sample size. The evidence room and the public ledger carry the full record. All insights.